The MIT-Broad Foundry: TA-2

Two critical bottlenecks in our capacity to rationally and predictably engineer biological systems are: (i) the limited number of well-characterized genetic elements from which to build (particularly for non-model, difficult organism hosts), and (ii) quality control of genetic constructs once built. Here, we will address these limitations by developing a rapid, high-throughput emulsion-based in vitro pipeline for prototyping genetic construct performance. We will identify robust, reliable genetic elements (e.g., promoter function, terminator strength, regulatory control, gene expression, etc.) prior to putting them in hosts. Uniquely, we will combine state-of-the-art cell-free protein synthesis (CFPS) systems, droplet based microfluidics, and novel methods in single droplet sequencing for measuring transcripts and proteins. The goals are two-fold. First, we will design, construct, confirm, and prototype function of all genetic parts prior to putting them in a host. In other words, we will use the emulsion system to know if each genetic construct is working as designed, at least with respect to the central dogma. Second, by combining multiple in vitro design-build-test-learn (DBT) iterations with relevant in vivo testing, we will use models to quantitatively map performance between these environments, for forward engineering . Our in vitro prototyping pipeline will allow substantial increases (>100x) in our ability to design biological systems. This unprecedented capability will come from the ability to (i) perform DBT iterations more rapidly than with cells (up to ~105 test constructs/conditions in individual droplets per hour, with costs as low as $0.02/droplet , (ii) avoid inherent limitations of time-consuming cloning and cultivation steps associated to living systems, (iii) debug and optimize genetic part activity through the use of simpler, well-defined experimental conditions and liquid automated handling, (iv) have a universal and impedance matched sequencing based readout leveraging high throughput sequencing efficiencies and scales where a single HiSeq 2500 flowcell could assay >1M construct/conditions from the droplet systems, and (v) focus characterization measurements on features reflective of genetic construct performance, which reduces cost via miniaturization, labor reduction, and scale for both RNA-seq and CFPS. In vitro prototyping approaches have been validated in E. coli CFPS (e.g., DNA regulatory elements, RNA genetic circuitry). Here, the key conceptual innovations are the integration of CFPS with microfluidics, as well as the development and utilization of novel CFPS systems from non-model organisms. Thus, while CFPS systems from E. coli can be integrated into droplet based microfluidics immediately, those from other target microorganisms (e.g., Streptomyces, chloroplasts, etc.) will be developed to achieve sufficient batch yields first (>50g/mL in a 3 hour batch reaction). It should not be overlooked that rapid and robust preparation of highly active cell extracts from genomically modified organisms will allow us to test genetic part operation under conditions conducive to the final host. For example, E. coli extracts could be made from strains harboring different mutations to observe potential context dependent effects. Once successful, the techniques that we propose will be integrated into the Foundry to offer quantitative improvements.